Colourfulness of the image and its application in image filtering

In this paper we show, from image processing point of view, different methods of computing of colourfulness of the image. We have calculated the colourfulness using simple estimate based on statistical parameters of the pixel cloud along red-green and yellow-blue axes. All experiments have been carried out on the set of natural colour images with different perceptual colourfulness. The relationships between colourfulness of the image and perceptual attributes (H, L, S) of pixels have been experimentally investigated. During image processing the colourfulness of the image can be changed but sometimes it should be preserved e.g. in image filtering. We have presented it on examples, which show that the difference of colourfulness can be useful for evaluating the image filtering algorithms beside such traditional performance functions as PSNR and DeltaE